P87. Expressive and receptive language mapping using ECoG and ECS

Abstract

Patients suffering from intractable epilepsy, who are candidates for surgical treatment, have to undergo several functional mapping procedures to make precisely diagnose for epileptogenic foci and eloquent brain regions. Since electrocorticography (ECoG) detects electric currents directly from the brain surface, it is expected to be the most reliable way to identify brain activity by focusing on changes of gamma-band oscillations. Previous studies identified the expressive and receptive language areas based on recordings with clinical standard ECoG grids ( Sinai et al., 2005 , Towle et al., 2008 ). In this study we investigated the sensitivity and specificity of passive ECoG mapping and the clinical standard electrical cortical stimulation (ECS) mapping using high-density electrode grids. The study included ECS and ECoG mapping with an epileptic patient having 236 subdural electrodes who underwent neuro-monitoring. The receptive and expressive language areas were each covered by a high-density subdural electrode grid with inter-electrode distance of 5 mm and conductive area of 1.5}mm. The ECoG mapping was performed with the real-time mapping system cortiQ (g.tec, Austria), which detects task related changes of gamma-band oscillations ( Prueckl et al., 2013 ). In order to activate the expressive language area, a picture naming task was performed during both mapping procedures. Additionally, a listening task was performed during the cortiQ mapping, to activate the receptive language area. The cortiQ mapping was repeated two times to ensure stable mapping results. The specificity and the sensitivity were computed for expressive and receptive language areas separately. Therefore, only the stimulated electrodes on the corresponding grid were used for comparison. During the ECS mapping the receptive language area was identified based on 24 tested sites over the left superior and middle temporal gyri. For expressive language mapping, 34 sites over the left inferior frontal gyrus were stimulated. The cortiQ mapping consistently showed activation patterns for both runs leading to a sensitivity of 50.0%/57.5% and a specificity of 85.0%/75.0% for the receptive/expressive language area. A next-neighbor approach that took into account the activation of surrounding electrodes showed a sensitivity of 100.0%/82.5% and a specificity of 85.0%/75.0% for receptive/expressive language mapping. The modalities show a good overlap, which is even more convincing due to the high number of tested sites in the area of interest. The passive ECoG mapping is a powerful supportive tool, as it allows reducing the number of tested sites during ECS.